Empiricism, Semantics and Ontology
1. The Problem of Abstract Entities
Basically the problem is, in the words of Quine "What is there?", or perhaps better more explicitly "What exists?", lest we should find a difference between the two.
2. Linguistic Frameworks
Carnap contends that there is an important distinction to be made between existence problems.
Some of them are, according to Carnap "internal questions" and others are "external questions".
Essential to this distinction is the notion of a "linguistic framework".
A linguistic framework is a set of linguistic conventions which determine the way in which we settle existence problems of a certain sort.
A straightforward example of such a linguistic system would be an axiomatised mathematical system, in which an existence problem is settled only by deduction from the axioms of the system.
The sort of existence problem thus settled would be what Carnap calls an internal question.
The "linguistic framework" itself, however, remains to be justified.
The question of the existence or otherwise of the total system of entities (e.g. of numbers in general, rather than of some particular number), which one might say is presupposed in the linguistic framework and by the asking and answering of internal questions, is Carnap's external question.
According to Carnap the internal questions are straightforward and philosophically uncontentious, and the external questions, which are generally those at issue in philosophical ontology disputes, are meaningless.
The external question should not be asked.
Instead we should ask whether or not any given linguistic framework is acceptable, and the answer to such a question should be based solely upon the utility or otherwise of accepting the framework.
3. What does acceptance of a kind of entities mean?
I think Carnap's answer to this amounts to "nothing".
It may be thought, he says, that prior to the acceptance of a new linguistic framework we must establish an affirmative answer to the external existence question.
However, for Carnap this question is devoid of meaning, and the only question to be settled is the utility or otherwise of the new framework.
4. Abstract entities in semantics.
In this section Carnap responds to criticism of his use of abstract entities as designata in accounts of the semantics of languages (which I take to be of his book "Meaning and Necessity" [Carnap47]) by Gilbert Ryle.
His defence is as you might expect.
He decouples use of abstract entities in semantics from any metaphysical belief in the existence of these entities.
This section is largely given over to a restatement of his general position on the meaninglessness of external questions with special reference to the use of abstract entities in semantics, and to a denial in this context of his having any beliefs or making any assertions of such ontological claims.
The following comments were originally written by me when preparing to write an undergraduate essay on ontology (see: Ontology) which turned out more sympathetic to Carnap's position than my immediate comments, which were:
Carnap's analysis is not entirely to my taste, but I think something of value can be obtained from it.
The main problem is that the analysis of existence questions into "internal" and "external" isn't quite satisfactory.
If our language is to be analysed into linguistic frameworks then it seems entirely plausible that these linguistic frameworks form a hierarchy, so that the external existence problem of one framework may be an internal existence question in another larger framework.
If we accept that these linguistic frameworks form a hierarchy, then it seems at least plausible that the totality of our language is a linguistic framework which encompasses all other frameworks.
In this case all existence questions except one become internal to some framework.
Anyone deliberating on whether frameworks "really" exist could retort to Carnap that he had accepted pragmatically the framework of our language as it is, and was now concerned to see whether this committed him to the existence of (for example) numbers.
The value which I do attach to Carnap's position is this.
He reduces the criterion for the existence problem of two, the linguistic and the pragmatic.
This seems to me satisfactory.
We must either settle existence problems by reference to preexisting linguistic conventions or we must produce a utilitarian justification of such linguistic innovation as we propose.
An appeal to "ontological insight" of some sort is not satisfactory.
No amount of argument can justify a usage which is netiher conventional nor useful.
I think I am today (2001-6-1), more in sympathy with Carnap than I was then.
I agree with Carnap in considering absolute questions about abstract ontology (i.e. his "external questions") as meaningless (except the question of consistency) and advocate a free-wheeling pluralistic pragmatism.
His distinction between internal and external questions seems to me important, and my injection of heirarchy unimportant.
In particular the possibility of settling external questions in some overall linguistic framework (discussed in my essay) now seems unattractive to me, and the question of ontological commitment in natural languages uninteresting.
created 1974-11-1 modified 2001-6-14